Lamina Cribrosa Depth Variation Measured by Spectral-Domain Optical Coherence Tomography Within and Between Four Glaucomatous Optic Disc Phenotypes

Abstract

Purpose: To study lamina cribrosa (LC) depth variation measured by spectral-domain optical coherence tomography (SD-OCT) in four glaucomatous optic disc phenotypes.

Methods: In this cross-sectional study, 59 normal eyes and 180 open-angle glaucoma (OAG) eyes were grouped into 56 focally injured discs (FI), 30 generalized enlargement of the optic cup discs (GE), 69 myopic glaucomatous discs (MY), and 25 senile sclerotic discs (SS). They were imaged by enhanced depth imaging SD-OCT, obtaining multiple horizontal and vertical optic disc B-scans. Mean and maximum LC depths were measured relative to Bruch's membrane opening (BMO) and the anterior sclera (AS) reference planes. Lamina cribrosa depths were compared between and among the normal and OAG group disc phenotypes. Lamina cribrosa depth differences within groups were examined as well as the overlap between them.

Results: Mean and maximum LC depths relative to the BMO and AS reference planes were greater in the OAG group than in the normal group (P < 0.0001). Among glaucomatous phenotypes, the GE group had the greatest (P < 0.001) and the SS group had the smallest (P < 0.05) mean and maximum LC depths. There was a wide range of LC depth overlap between the normal and SS groups, and a high proportion of SS eyes had LC depths within the 95% confidence interval of the normal group.

Conclusions: The LC was displaced posteriorly in OAG group compared to the normal group. The LC depth was significantly different among four glaucomatous disc phenotypes. The LC depth of the SS group was similar to the normal group.

The lamina cribrosa (LC) is recognized as the principal site of retinal ganglion cell axonal injury in glaucoma.^1,2 Deformation of the LC may promote glaucomatous optic neuropathy by blocking the axoplasmic flow within the optic nerve fibers.³ In the past the LC morphology has been studied largely by histologic methods.^1–10 Recent invention of spectral-domain optical coherence tomography (SD-OCT)¹¹ and enhanced depth imaging (EDI)^12,13 has enabled the visualization of deep optic disc structures in living human eyes.¹⁴ It is now possible to observe the anterior surface of the LC and evaluate the LC position by measuring the distance from a reference plane to the anterior LC surface. 

The LC position of glaucoma eyes has been investigated in numerous studies. Posterior displacement of the LC has been reported in glaucoma eyes compared to the normal eyes,¹⁵ and the degree of posterior displacement is greater in the eyes with visual field (VF) defects than in fellow eyes without VF defects.¹⁵ Anterior displacement of the LC has been reported after surgical^16–19 and medical²⁰ reduction of the intraocular pressure (IOP). The amount of anterior LC displacement is associated with the amount of IOP reduction,^16,18–20 but this association has not been found in another study.¹⁷ The influence of LC position and LC thickness on the rate of retinal nerve fiber layer thinning has been studied in open-angle glaucoma (OAG).²¹ The authors report that greater posterior LC displacement and thinner LCs are associated with a faster progression of retinal nerve fiber layer thinning. 

Glaucoma is considered to be a multifactorial disease, with various risk factors affecting its development. Some of these factors may determine the appearance of the damaged glaucomatous optic discs. In the past, distinct appearances of the glaucomatous optic discs have been described, each of which is thought to correlate to specific risk factors.^22–26 We hypothesized that each of the distinctly different optic disc appearances may exhibit a uniquely different LC morphology. To our knowledge, no earlier studies have investigated the LC features, including the LC position, in the different glaucomatous optic disc phenotypes. 

In this study, we classified OAG eyes into four phenotypes according to the optic disc appearance and assessed their LC position by measuring the anterior LC depth. We examined the differences in LC depth within the four glaucomatous optic disc phenotypes as well as the overlap among them. 

Methods

This study was approved by the Institutional Review Board of Akita University Graduate School of Medicine. It was performed with the written informed consent of the participants and followed the tenets of the Declaration of Helsinki. 

Participants

Patients with OAG were recruited from the outpatient clinic of Akita University Graduate School of Medicine from September 2012 to July 2013. Normal subjects were recruited by advertisement within the Akita University Hospital. All subjects were Japanese. 

Each subject underwent a comprehensive ophthalmic assessment, including refraction tests, measurement of best-corrected visual acuity, slit lamp biomicroscopy, gonioscopy, stereoscopic examination of the optic discs, and color fundus stereo photography (Canon, Tokyo, Japan). In the pseudophakic eyes, the refraction error before surgery was used. Intraocular pressure was measured by Goldmann applanation tonometry. The average IOP, obtained in two to three measurements before the initiation of therapy, was analyzed as well as the IOP on the day of OCT imaging. The Humphrey VF test was performed by using the 30-2 Swedish Interactive Threshold Algorithm standard program (Carl Zeiss Meditec, Dublin, CA, USA). 

To be included, eyes had to have a best-corrected visual acuity of ≥20/30 to minimize the effect of media opacity. Excluded eyes were those with intraocular diseases, previous intraocular surgery except uncomplicated cataract extraction and glaucoma surgery, ocular trauma, myopic macular changes such as patchy chorioretinal atrophy, lacquer crack lesions, choroidal neovascularization, or neurologic diseases that might affect the optic disc appearance or the VF. 

A diagnosis of OAG was made when the patient had an open-angle, glaucomatous optic disc change (localized or diffuse rim thinning or retinal nerve fiber defect), and glaucomatous VF defects corresponding to the optic disc changes. The glaucomatous VF defects included glaucoma hemifield test results outside normal limits, or the presence of at least three contiguous nonedge test points within the same hemifield on the pattern deviation plot at <5%, with at least one of these points at <1%, which was confirmed on two consecutive tests. Normal eyes were required to have a normal-appearing angle, IOP ≤ 21 mm Hg, clinically normal optic discs, and no VF defect. 

Glaucomatous optic disc appearance was classified by reviewing stereo photographs. They were classified into one of the following four phenotypes^24,25 (Fig. 1): focally injured discs (FI; Fig. 1A) that were previously referred to as focal ischemic discs, generalized enlargement of the optic cup discs (GE; Fig. 1B), myopic glaucomatous discs (MY; Fig. 1C), and senile sclerotic discs (SS; Fig. 1D). The classification was performed independently by two glaucoma specialists in a masked fashion. Eyes with a mixture of several disc appearances were excluded. If the observers did not agree on the classification, they attempted to reach a consensus. If they could not, then the eye was excluded. 

Figure 1

 

Classification of the glaucomatous optic discs. (A) Focally injured discs with localized tissue loss at the superior or inferior pole and with relatively intact neuroretinal rim in the other areas of the disc. (B) Generalized enlargement of the optic cup discs with diffusely enlarged, round cups without a localized defect of the neuroretinal rim. (C) Myopic glaucomatous discs with tilted discs and peripapillary atrophy. Additional evidence of glaucomatous damage included thinning of the superior and inferior rim. (D) Senile sclerotic discs with a saucer-shaped and shallow cup with peripapillary atrophy.

EDI-OCT of the Optic Discs

Enhanced depth imaging–OCT cross-sectional B-scans of the optic discs were performed with a Spectralis OCT system (Heidelberg Engineering, Heidelberg, Germany). Images were obtained in a 10 × 20-degree rectangle centered on the optic disc both vertically and horizontally (Figs. 2A–D). Approximately 70 B-scan images, 30 μm apart, covering the optic discs were obtained in each scan direction. For each B-scan image, 25 frames were averaged. The scan images obtained at the central two-thirds of the optic discs were assessed because the image quality of this region was generally good, with less influence of the overlying structures such as retinal vessel trunks and the neuroretinal rim. This region was divided into six equal parts by five B-scan images, and the LC depths of these five images were measured (Figs. 2A, 2B). If the quality of the selected scan image was suboptimal, the neighboring image was assessed. If three consecutive images were suboptimal and unusable, the eye was excluded. 

Figure 2

 

(A, B) Optic disc photographs presenting the location of LC depth measurement. The B-scans obtained at the central two-thirds of the optic discs were assessed, because the image quality of this region was generally good with less influence of the overlying structures such as retinal vessels and neuroretinal rim. This region was divided into six equal parts by five B-scan images, and the LC depths of these five images were assessed. The depth of the five horizontal and five vertical images, totaling 10 B-scan images, was averaged and defined as the LC depth of the eye. (C) 3D volumetric image of an optic disc reconstructed from scan data set obtained in 10 × 20-degree rectangle centered on the optic discs. Orange line indicates the location of one of the selected five B-scan lines. (D) The EDI-OCT image obtained at the location of the orange B-scan line. (E) The EDI-OCT scan image of the optic nerve head. (F) The same image with the labels. Two reference lines were used for the LC depth measurement: BMO line, which connects termination points of Bruch's membrane (blue line), and AS line, which connects two points of the anterior scleral surface located at 1750 μm from the center of the BMO (red line). The definition of the AS line was based on the modified method of Johnstone et al.²⁷ This location was chosen because the anterior scleral surface was most visible in this location when viewing the scan data, while its visibility was often poor in the regions more adjacent to the neural canal owing to the shadowing from the overlying structures. The anterior surface of the LC was manually drawn (white dotted line), and the length between the reference lines and the anterior LC surface was defined as the LC depth (lines with arrows).

The LC depth was measured by using two reference planes: Bruch's membrane opening (BMO) and the anterior sclera (AS; Figs. 2E, 2F). The BMO reference line was drawn by connecting the two termination points of Bruch's membrane on each B-scan.^15–21 The AS reference line was defined as the line connecting two points of the anterior scleral surface located 1750 μm from the center of the BMO in each B-scan. The definition of the AS line was based on the modified method of Johnstone et al.²⁷ This location was chosen because the scleral surface was most visible in this location when viewing the scan data, while its visibility was often poor in the regions more adjacent to the neural canal owing to the shadowing from the overlying structures.^27,28 The LC depth was defined as the length of a line perpendicular from these reference lines to the manually drawn anterior surface of the LC (Fig. 2F). The LC depth was measured at the center of the reference line and two middle points between the center and the termination points of the reference line. The depth of these three points was averaged and defined as the LC depth of the B-scan image. Then, the depth of the five horizontal and five vertical B-scan images was averaged and defined as the LC depth of the eye. The B-scan with the largest LC depth was chosen and defined as the maximum LC depth of the eye. The mean and maximum LC depths relative to the BMO and AS reference planes were compared between normal and OAG eyes, and then among the normal and the four optic disc phenotypes. To examine the overlap of the LC depth between the normal and the four glaucomatous phenotypes, the number of eyes in which the LC depth fell within the 95% confidence interval (CI) of the other disc types was counted. 

Statistical Analysis

The difference between normal and OAG eyes was analyzed by Student's t-test. To adjust for the possible covariates, analysis of covariance (ANCOVA) was performed. One-way analysis of variance (ANOVA) was used to compare data among the normal and the four disc phenotypes. When the F test was statistically significant, the Scheffe multiple comparison method was used. The comparison of the proportion among groups was performed by using 2 × 4 χ² test. The level of significance was set at P < 0.05. All analyses, with two-sided P values, were performed by using SPBS version 9.66.²⁹ 

Reproducibility of the LC depth measurements was assessed in 10 randomly selected normal eyes. Analysis was based on three independent series of evaluations made by two independent observers. Using a two-way mixed effect model, the absolute agreement of a single observer's measurement was calculated as the intraclass correlation coefficient (ICC 1). The agreement of the mean of three measurements by the two observers was calculated as the interclass correlation coefficient (ICC 2). 

Results

In 420 eyes of 210 OAG patients, eyes were excluded for the following reasons: poor quality EDI-OCT images (n = 64), difficulty in classifying optic disc appearance into one phenotype (n = 56), and difficulty in measuring LC depth in a reproducible manner owing to a downward LC insertion (n = 3). Both eyes were eligible in 107 patients, and one eye for each patient was randomly selected. One eye was eligible in 73 patients. Thus, 180 OAG eyes were included. Of the 66 normal subjects recruited, 59 eyes were included as controls. 

Agreement on the masked evaluation of the optic disc photographs in the diagnosis of normal and glaucoma eyes between observers was 0.962 (95% CI = 0.930–0.983). Agreement on the categorization of the glaucomatous disc appearances between observers was 0.969 (95% CI = 0.943–0.985). The reproducibility of the LC depth measurements was determined. For the mean LC depth relative to the BMO, the ICC 1 was 0.988 for observer 1 and 0.992 for observer 2, and the ICC 2 was 0.976 with a 95% CI of 0.962 to 0.988. For the maximum LC depth relative to the BMO, the ICC 1 was 0.990 for observer 1 and 0.987 for observer 2, and the ICC 2 was 0.978 with a 95% CI of 0.959 to 0.994. For the mean LC depth relative to the AS, the ICC 1 was 0.975 for observer 1 and 0.981 for observer 2, and the ICC 2 was 0.974 with a 95% CI of 0.959 to 0.990. For the maximum LC depth relative to the AS, the ICC 1 was 0.978 for observer 1 and 0.977 for observer 2, and the ICC 2 was 0.972 with a 95% CI of 0.959 to 0.986. The P values for all of the reproducibility data were <0.001. 

There were no differences in sex proportion or age between the normal and the OAG group, while the refractive error was greater for the OAG group (P < 0.0001, Table 1). There were no differences in IOP among the groups on the day of OCT imaging. Mean deviation (MD) of the Humphrey VF test of OAG eyes was −8.38 ± 7.16 dB. The mean and maximum LC depths relative to both the BMO and the AS reference planes were significantly greater in the OAG group than in the normal group (all P < 0.0001, Table 1). The statistical significance remained after adjusting for sex, age, refractive error, IOP, and MD. 

Table 1

 

Clinical Characteristics and Lamina Cribrosa Depths of Normal and Open-Angle Glaucoma Eyes

 

Among 180 OAG eyes, the optic discs of 56 were classified as FI, 30 as GE, 69 as MY, and 25 as SS (Table 1). The sex proportion did not differ among phenotypes, though there was a significant difference in age. Patients of the FI group were older than those of the MY group (P < 0.0001), and patients of the SS group were older than those of the GE and MY groups (P = 0.0074 and P < 0.0001, respectively). The MY group had a greater refractive error than the other disc groups (all P < 0.0001). There were significant differences in the IOP among groups both before treatment and on the day of OCT imaging (P < 0.0001 and P = 0.0024, respectively; Table 1). The GE group had higher IOP before treatment than the other disc groups (all P < 0.0001). The IOP on the day of OCT imaging was higher in the GE group than in the MY group (P = 0.0031), while it was not different from that in the FI and SS groups. The VF damage, evaluated as the MD of the Humphrey VF test, did not differ among phenotypes. There were significant differences in mean and maximum LC depths relative to both the BMO and AS reference planes within phenotypes (all P < 0.0001, Table 1). The significance remained after adjusting for sex, age, refractive error, IOP, and MD. 

Comparison of the LC depth relative to the AS reference plane among the normal and four glaucoma optic disc phenotypes showed that all of the glaucoma disc groups had greater mean and maximum LC depths than did the normal group, except for the SS group (Fig. 3). There was no significant difference in the LC depth between the SS group and the normal group. Among the four glaucomatous phenotypes, the GE group had greater mean and maximum LC depths than other phenotypes (all P < 0.001, Fig. 3), while the SS group had a smaller depth than others (all P < 0.05, Fig. 3). The results of the LC depth comparison among groups, using the BMO reference, were compatible with those with the AS reference (data not shown). 

Figure 3

 

Comparison of the LC depth among the normal and four glaucomatous optic disc phenotypes. Scheffe multiple comparison method revealed significant differences among groups. The LC depths were measured relative to the anterior scleral reference plane. The depth values were adjusted for sex, age, refractive error, IOP, and MD.

There was a wide range of overlap of the LC depth between the normal group and the SS group (Table 2). The proportion of the eyes in which the mean and maximum LC depths fell into the 95% CI of normal group was higher in the SS group than in the FI and GE group (24.0% for SS, 5.4% for FI, and 0% for GE in the mean LC depth; and 20.0% for SS, 1.8% for FI, and 0% for GE in the maximum LC depth; all P < 0.05; Table 2). The GE group had little overlap with other disc groups because of its greater LC depth. In only one GE eye (3.3%) did the mean and maximum LC depths overlap with the FI group (Table 2). 

Table 2

 

Number of Eyes in Each Optic Disc Phenotype Falling Within the 95% CI of the Other Disc Phenotypes

 

Discussion

In the current study we demonstrated significant differences in the LC depths within four glaucomatous optic disc phenotypes. The GE group had greater and SS group had smaller LC depths than other disc groups. There was a wide range of overlap of the LC depths between the normal group and SS group, and a high proportion of eyes of the SS group had LC depths that were within the 95% CI of that of the normal group. 

In the past, many studies have investigated morphologic changes of the LC in glaucomatous eyes. Various findings have been reported, including initial thickening,^30,31 subsequent thinning,² posterior displacement,^1,2,15 posterior migration of the laminar insertion,^32,33 and partial or total thickness disruption.^34–38 Several factors have been reported to be associated with the status and the magnitude of the LC deformation. One is the disease stage, in which greater LC displacement is observed in the eyes with more advanced stages of glaucoma.^2,39 Another is aging,^40–43 during which older eyes present less LC deformation than the younger eyes for the same level of functional loss.^40–42 

In the current study, we measured LC depth relative to the AS as well as the BMO. Previous studies have used BMO as a reference plane. However, recent cross-sectional studies have demonstrated significant thinning of the choroid and the posterior migration of the BMO with aging.^44,45 This implies that the BMO is not a stable reference. Alternatively, the AS has been proposed as a reference that could eliminate the influence of choroidal thickness.^27,28 With either reference plane, our results consistently demonstrated that the LC depth was greater in glaucoma eyes than in normal eyes, and there were significant depth variations among the four optic disc phenotypes. 

Previous studies using EDI-OCT with the BMO reference plane have reported the LC depths of normal and glaucoma eyes. Furlanetto et al.¹⁵ have found that the mean and maximum LC depths of normal eyes are 355 and 438 μm, and those of glaucoma eyes are 453 and 570 μm. Seo et al.⁴⁶ have reported mean and maximum LC depths of normal eyes as 402 and 425 μm. We found that the mean and maximum LC depths of normal eyes were 388 and 405 μm, and in glaucoma eyes they were 492 and 520 μm. Our results are compatible with both studies,^15,46 and demonstrated posterior displacement of the LCs in most glaucoma eyes. 

In the current study, the GE disc group presented significantly greater LC depth than the other disc groups. Patients with GE discs were younger than those of the other disc groups except for those with MY discs, and they had higher IOP on the day of imaging. When the damage was created, the patients were younger and the laminar tissue might have been more compliant. Therefore, they manifested more deformation as part of the damage before the institution of IOP-lowering therapy. On the day of imaging, the IOP of the GE disc groups was within normal range; however, it was still higher than that of other disc groups. The posterior deformation of the LC in glaucoma consists of permanent and reversible components, and the reversible component may be determined by the IOP level.^16–20,47 The fact that the IOP was higher in the GE disc group than in the others on the day of imaging indicates that the reversible component might have been added to their LC deformation, which would not be present if all eyes were imaged at the same, sufficiently low level of IOP. 

The patients of the SS group were older and had shallower LC depths than those in the other disc groups in the current study. The appearance of the SS disc has been recognized as a characteristic of aged glaucomatous optic disc.^22,24,25 Recent studies using SD-OCT have reported that LC depths are shallower in older eyes than in younger eyes for the same level of glaucomatous VF defect.^41,42 Our study is in line with these studies in finding shallower LC depth in the SS group than in the other disc groups with the same level of VF defect. It is noteworthy that the LC depth of the SS group had considerable overlap with that of the normal group. It suggests that retinal ganglion cell axon loss leading to a glaucomatous pattern of VF damage can occur in SS discs by mechanisms that do not necessarily result in obvious posterior LC deformation. 

There are several limitations in the current study. First, the OCT data were not acquired relative to the foveal–BMO axis,^48–50 because some of our early data were taken before the widespread acknowledgement of the concept. We measured LC depth in both horizontal and vertical scans, which allowed us to evaluate the LC position; however, ideally, it should be done relative to the foveal–BMO axis to provide anatomic consistency. Second, the LC depth was measured at three points in one scan line. Theoretically, fewer measurement points are prone to cause errors, and more points increase the accuracy of the results. Because the identification of the anterior LC surface was sometimes difficult in certain areas of the images, it was infeasible to increase the measurement points in all scan lines. However, because intraobserver and interobserver agreements of the depth measurements were excellent, we believe the reliability of our data is acceptable for the analysis. Third, the identification of the anterior scleral surface was difficult in some eyes; therefore, the position of the AS reference might not be totally accurate. However, because the significance levels of the depth differences among the groups were high, and the results were compatible to the measurements based on the BMO reference plane, this limitation is not likely to affect the main outcome of our study. Finally, all of the subjects were Japanese; therefore, the results might not be totally applicable to other ethnic groups. Further studies are needed in other ethnic groups to see the compatibility of the results. 

In conclusion, we described for the first time the variation of LC position within four glaucomatous optic disc phenotypes. We urge clinicians to be cognizant of this variation in studying morphologic changes of the LCs in glaucoma eyes. Further studies are needed to investigate the significance of these findings in the association with functional loss. 

Acknowledgments

The authors alone are responsible for the content and writing of the paper. 

Disclosure: Y. Sawada, None; M. Hangai, None; K. Murata, None; M. Ishikawa, None; T. Yoshitomi, None